In October 2000, the GAIA astrometric mission wasapproved as one ofthe next two "cornerstones" of ESA's science programme, with a launchdate target of 2010-12. GAIA will provide positional and radialvelocity measurements with the accuracies needed to produce astereoscopic and kinematic census of about one billion starsthroughout our Galaxy (and into the Local Group), amounting to about1 per cent of the Galactic stellar population. GAIA's main scientificgoal is to clarify the origin and history of our Galaxy, from aquantitative census of the stellar populations. It will advancequestions such as when the stars in our Galaxy formed, when and how itwas assembled, and its distribution of dark matter. The survey aimsfor completeness to V=20 mag, with accuracies of 10 μas at15 mag.

Three topics within GAIA photometry have been selected for discussionhere. (1) The sampling scheme of the CCDs has been tuned to serve thevarious purposes: detection of stars, astrometry, broad- andmedium-band photometry, and imaging of a small area around eachdetected star, but the tuning may not yet be optimal. (2) Thecalibration of CCD sensitivity is critical since we aim formillimagnitude accuracy, the spikes of the diffraction image aresuited for an accurate calibration of the magnitude scale, and forextending the range of astrometry and photometry to very bright stars.(3) Finally the various photometric filter systems are illustrated.

GAIA spectroscopic and radial velocity performances are reviewed onthe base of ground-based test observations and simulations. Theprospects for accurate analysis of stellar atmospheres (temperature,gravity, chemical abundances, rotation, peculiarities) and preciseradial velocities (single stars, binaries, pulsating stars) arecolourful provided the spectral dispersion is high enough. A higherdispersions also favors a given precision of radial velocities to bereached at fainter magnitudes: for example, with current parametersfor GAIA spectrograph, a 1 km s-1 accuracy on individual epochRVs of a K0 star is reached at V~13.0 mag with 0.25 Å/pixdispersion spectra, at V~10.3 mag for 0.5 Å/pix, andV~6.7 mag for 1 Å/pix. GAIA radial velocities for singlestars can match the ~0.5km s-1 mean accuracy oftangential motions at V=15 mag, provided the observations areperformed at a dispersion not less than 0.5 Å/pix.

GAIA will obtain, in a five-year mission, astrometric, photometricand spectroscopic data for more than a billion stars. The treatment ofthis huge amount of data is a challenge for the astronomicalcommunity, not only due to the database size, but also to the complexrelationships existing between the different sets of data and with theinstrumental and orbital parameters. The GAIA Data Analysis Study(GDAAS), under ESA's contract, aims to prove the feasibility of thedata reduction scheme and to implement and test a prototype of such alarge database. We present the main aspects that GDAAS addresses andreview its present status.

The successful hipparcos missionhas brought, and proved the validity of, a new observationalconcept to achieve absolute measurements of star positions inspace. Only five years after the final publication of the resultsthere are now four astrometry missions scheduled for the nextdecade, three of which appearing as the natural heirs of thispioneering mission. Space astrometry missions share a certainnumber of common features imposed by their observationalprinciples and their scientific goals. The first part of thispaper attempts to show how the objectives put severe constraintson the design and that this can be investigated with a fairlygeneral approach. The German mission diva is then detailed as anillustration (gaia being considered elsewhere in this volume).In the last section few words are added about fame and sim followed by global comparisons of these missions and theirscientific returns.

The problem of relativistic modelling of GAIA observations isreviewed. Basic principles of the relativistic model and its overallstructure are described in detail. All relativistic effects which mayamount to 1 μas are revealed and discussed. It is argued thatrelativistic definitions of astrometric parameters (position,parallax, proper motion, radial velocity, etc.) consistent with anaccuracy of 1 μas should be considered only within a well-definedalgorithm of relativistic reduction of observational data.

The baseline of GAIA is the study of the formation and evolution ofthe Milky Way from astrometric and photometric measurements. Themission displays a remarkable versatility in its applications,including a significant impact in fundamental physics. In thefollowing I will try to define the relationship between astrometry andfundamental physics, before considering in more detail thedetermination of the space curvature and the study of the nonlinearity of gravity from the astrometric measurements. It is shownthat GAIA is perfectly at home to sense the bending of light-rays inthe solar gravitational field and should be able to determine the PPNparameter γ with a precision between 10-6 and 10-7,much better than any other determination expected by 2015. Afavorable combination of distances and eccentricities on a handful ofminor planets will permit to search for the relativistic perihelionprecession. The parameter β should be ascertained to 10-3-10-4 provided the solar quadrupole moment is not solvedsimultaneously and constrained from other sources. This achievableaccuracy rests also on assumptions regarding the reconstruction oforbits of minor planets from GAIA-only observations.

In this paper we examine the effects introduced in the luminosityfunction of white dwarf stars by a change of the gravitationalconstant during the life of the Universe. A decrease of G inducesthe expansion of the star and accelerates its cooling. This effecthas noticeable effects on the luminosity function and can be used toset upper bounds on the rate of change of G. Preliminary resultsindicate that Ġ/G ≤ 10-13 yr-1. These values are thebest ever obtained and will improve as the white dwarf luminosityfunction will improve.

Understanding the stellar interiors is an important issue forastrophysics, but also for fundamental physics. The physicaldescription of the stellar material implies statistical physics aswell as hydrodynamics. We give examples of unsolved physical problemsand their consequences for predictions of stellar properties andevolution. It is shown that the accumulation of high-quality datalike distances and photometry on a large variety of objects, asproposed by GAIA, will be very useful. During the same period,asteroseismology will develop both from the ground in velocimetry, andfrom space in photometry, providing for the closest stars, a directinsight into their interiors. This ensemble of new data willconstitute the basis of the general physical description of stellarinteriors.

We show that there is a wide range of problems, where high accuracy astrometry can bring new information in stellar evolution, not limited to the classicalapplications in the HR diagram, such as thedeterminations of the helium and metal contents and of the size of the convective core.In particular, we show some new applications to the study of star formation. The accuratemeasurement of the luminosity of the observed birthline in the HR diagram is a key information very sensitiveto the mass accretion rates. These rates determine all the properties of the pre-MS models. For the evolution of massivestars with mass loss and rotational mixing, strong constraints can be obtained from the combination of astrometric data and abundance determinations obtained by highresolution spectroscopy. We also emphasize that a better knowledge of the luminosity distribution of WR stars is strongly constraining the final masses at the time of supernovaexplosions as well as the nucleosynthetic yields.

The study of the oldest stellar populations in the Galaxy providesimportant cosmological information (e.g. a lower limit to the age ofthe Universe) and helps constraining models of galaxy formation. Themost commonly used stellar age indicators are reviewed and discussed,and the impact of GAIA on these topics is estimated.

Stellar candles are the basic tools to determine the extragalacticdistance scale. Its use relies on some given fundamentalrelation that allows to estimate the absolute magnitude of an objectfrom observable parameters and from that its distance. The Hipparcosmission greatly contributed to the refinement of these fundamentalrelations, but even with its milliarcsecond-precision data, theuncertainties in the calibration of Stellar Candles are still high. Toimprove these results larger samples and higher astrometric precisionare needed. GAIA, with its microarcsecond, one-billion objectscatalogue, will provide all the necessary data for this task.

We present, in a first part,a short review of existing andplanned large catalogues and surveys.In a second part, we present the current plansfor building the Virtual Observatories of the future. Finally we givea few examples ofemerging methodologies for the analysisof huge catalogues.

Important progress has been made within the past few years regardingthe theory of low mass stars (m < 1 M⊙) and brown dwarfs(m < 0.075 M⊙).The main improvements concern the equation of state of dense plasmasand the modelling of cool and dense atmospheres,necessary for a correct description of such objects. Thesetheoretical efforts now yield a better understanding of these objectsand good agreement with observations regarding colour-magnitudediagrams of globular clusters, mass-magnitude relationshipsand near-IR colour-magnitude diagrams for openclusters. However uncertainties still remain regardingsynthetic optical colours and the complex problemof dust formation in the coolest atmosphere models.We will present a review of the current success and uncertaintiesof the theory of these objects, which may guide forthcoming projectswith GAIA.

Because of their very red colours and intrinsic faintness, field browndwarfs will represent a small but valuable subset of the GAIAcatalogue. The return of the astrometric satellite is expected to beimportant because of the inherent difficulty of obtaining goodparallaxes in general and for this class of objects in particular. Ourfirst estimates show that, due to the photometric sensitivity of theastrometric CCD (ASM1) towards relatively blue objects, GAIA isunlikely to detect field brown dwarfs that have not been already seenis previous near-IR surveys, to the notable exception of the galacticplane region. The real advantage of GAIA over ground-based surveyswill be the very accurate (to within a few percents) astrometric datafor a few thousands brown dwarfs. These data should permit a detailedmapping of the transition region between stellar and substellarregimes, together with the kinematical and density patterns of theyoungest brown dwarfs in our neighbourhood.

The ESA Cornerstone Mission GAIA, to be launched prior to 2012 and witha nominal lifetime of 5 years, will improve the accuracy of Hipparcosastrometry by more than two orders of magnitude.GAIA high-precision global astrometric measurements will provide deep insightson the science of extra-solar planets. The GAIA contribution is primarilyunderstood in terms of the number and spectral type of targets availablefor investigation, and characteristics of the planets to be searched for.Several hundreds of thousands of solar-type stars (F-G-K) within a sphereof ~200 pc centered on our Sun will be observed. GAIA will be particularlysensitive to giant planets (MP ~ MJ)on wide orbits, up to periods twice as large as the mission duration, thepotential signposts of the existence of rocky planets in the Habitable Zone.Thousands of new planets might be discovered, and a significant fractionof those which will be detected will have orbital parameters measured tobetter than 30% accuracy. By measuring to a few degreesthe relative inclinations of planets in multiple systems with favorableconfigurations, GAIA will also make measurements of unique value towards abetter understanding of the formation and evolution processes of planetarysystems.